Use a TensorFlow Lite model for inference with ML Kit on iOS

You can use ML Kit to perform on-device inference with a TensorFlow Lite model.

ML Kit can use TensorFlow Lite models only on devices running iOS 9 and newer.

Before you begin

  1. If you have not already added Firebase to your app, do so by following the steps in the getting started guide.
  2. Include the ML Kit libraries in your Podfile: 
    pod 'Firebase/MLModelInterpreter', '6.25.0'
    After you install or update your project's Pods, be sure to open your Xcode project using its .xcworkspace.
  3. In your app, import Firebase:

    Swift

    import Firebase

    Objective-C

    @import Firebase;
  4. Convert the TensorFlow model you want to use to TensorFlow Lite format. See TOCO: TensorFlow Lite Optimizing Converter.

Host or bundle your model

Before you can use a TensorFlow Lite model for inference in your app, you must make the model available to ML Kit. ML Kit can use TensorFlow Lite models hosted remotely using Firebase, bundled with the app binary, or both.

By hosting a model on Firebase, you can update the model without releasing a new app version, and you can use Remote Config and A/B Testing to dynamically serve different models to different sets of users.

If you choose to only provide the model by hosting it with Firebase, and not bundle it with your app, you can reduce the initial download size of your app. Keep in mind, though, that if the model is not bundled with your app, any model-related functionality will not be available until your app downloads the model for the first time.

By bundling your model with your app, you can ensure your app's ML features still work when the Firebase-hosted model isn't available.

Host models on Firebase

To host your TensorFlow Lite model on Firebase:

  1. In the ML Kit section of the Firebase console, click the Custom tab.
  2. Click Add custom model (or Add another model).
  3. Specify a name that will be used to identify your model in your Firebase project, then upload the TensorFlow Lite model file (usually ending in .tflite or .lite).

After you add a custom model to your Firebase project, you can reference the model in your apps using the name you specified. At any time, you can upload a new TensorFlow Lite model, and your app will download the new model and start using it when the app next restarts. You can define the device conditions required for your app to attempt to update the model (see below).

Bundle models with an app

To bundle your TensorFlow Lite model with your app, add the model file (usually ending in .tflite or .lite) to your Xcode project, taking care to select Copy bundle resources when you do so. The model file will be included in the app bundle and available to ML Kit.

Load the model

To use your TensorFlow Lite model in your app, first configure ML Kit with the locations where your model is available: remotely using Firebase, in local storage, or both. If you specify both a local and remote model, you can use the remote model if it is available, and fall back to the locally-stored model if the remote model isn't available.

Configure a Firebase-hosted model

If you hosted your model with Firebase, create a CustomRemoteModel object, specifying the name you assigned the model when you published it:

Swift

let remoteModel = CustomRemoteModel(
  name: "your_remote_model"  // The name you assigned in the Firebase console.
)

Objective-C

// Initialize using the name you assigned in the Firebase console.
FIRCustomRemoteModel *remoteModel =
    [[FIRCustomRemoteModel alloc] initWithName:@"your_remote_model"];

Then, start the model download task, specifying the conditions under which you want to allow downloading. If the model isn't on the device, or if a newer version of the model is available, the task will asynchronously download the model from Firebase:

Swift

let downloadConditions = ModelDownloadConditions(
  allowsCellularAccess: true,
  allowsBackgroundDownloading: true
)

let downloadProgress = ModelManager.modelManager().download(
  remoteModel,
  conditions: downloadConditions
)

Objective-C

FIRModelDownloadConditions *downloadConditions =
    [[FIRModelDownloadConditions alloc] initWithAllowsCellularAccess:YES
                                         allowsBackgroundDownloading:YES];

NSProgress *downloadProgress =
    [[FIRModelManager modelManager] downloadRemoteModel:remoteModel
                                             conditions:downloadConditions];

Many apps start the download task in their initialization code, but you can do so at any point before you need to use the model.

Configure a local model

If you bundled the model with your app, create a CustomLocalModel object, specifying the filename of the TensorFlow Lite model:

Swift

guard let modelPath = Bundle.main.path(
  forResource: "your_model",
  ofType: "tflite",
  inDirectory: "your_model_directory"
) else { /* Handle error. */ }
let localModel = CustomLocalModel(modelPath: modelPath)

Objective-C

NSString *modelPath = [NSBundle.mainBundle pathForResource:@"your_model"
                                                    ofType:@"tflite"
                                               inDirectory:@"your_model_directory"];
FIRCustomLocalModel *localModel =
    [[FIRCustomLocalModel alloc] initWithModelPath:modelPath];

Create an interpreter from your model

After you configure your model sources, create a ModelInterpreter object from one of them.

If you only have a locally-bundled model, just pass the CustomLocalModel object to modelInterpreter(localModel:):

Swift

let interpreter = ModelInterpreter.modelInterpreter(localModel: localModel)

Objective-C

FIRModelInterpreter *interpreter =
    [FIRModelInterpreter modelInterpreterForLocalModel:localModel];

If you have a remotely-hosted model, you will have to check that it has been downloaded before you run it. You can check the status of the model download task using the model manager's isModelDownloaded(remoteModel:) method.

Although you only have to confirm this before running the interpreter, if you have both a remotely-hosted model and a locally-bundled model, it might make sense to perform this check when instantiating the ModelInterpreter: create an interpreter from the remote model if it's been downloaded, and from the local model otherwise.

Swift

var interpreter: ModelInterpreter
if ModelManager.modelManager().isModelDownloaded(remoteModel) {
  interpreter = ModelInterpreter.modelInterpreter(remoteModel: remoteModel)
} else {
  interpreter = ModelInterpreter.modelInterpreter(localModel: localModel)
}

Objective-C

FIRModelInterpreter *interpreter;
if ([[FIRModelManager modelManager] isModelDownloaded:remoteModel]) {
  interpreter = [FIRModelInterpreter modelInterpreterForRemoteModel:remoteModel];
} else {
  interpreter = [FIRModelInterpreter modelInterpreterForLocalModel:localModel];
}

If you only have a remotely-hosted model, you should disable model-related functionality—for example, gray-out or hide part of your UI—until you confirm the model has been downloaded.

You can get the model download status by attaching observers to the default Notification Center. Be sure to use a weak reference to self in the observer block, since downloads can take some time, and the originating object can be freed by the time the download finishes. For example:

Swift

NotificationCenter.default.addObserver(
    forName: .firebaseMLModelDownloadDidSucceed,
    object: nil,
    queue: nil
) { [weak self] notification in
    guard let strongSelf = self,
        let userInfo = notification.userInfo,
        let model = userInfo[ModelDownloadUserInfoKey.remoteModel.rawValue]
            as? RemoteModel,
        model.name == "your_remote_model"
        else { return }
    // The model was downloaded and is available on the device
}

NotificationCenter.default.addObserver(
    forName: .firebaseMLModelDownloadDidFail,
    object: nil,
    queue: nil
) { [weak self] notification in
    guard let strongSelf = self,
        let userInfo = notification.userInfo,
        let model = userInfo[ModelDownloadUserInfoKey.remoteModel.rawValue]
            as? RemoteModel
        else { return }
    let error = userInfo[ModelDownloadUserInfoKey.error.rawValue]
    // ...
}

Objective-C

__weak typeof(self) weakSelf = self;

[NSNotificationCenter.defaultCenter
    addObserverForName:FIRModelDownloadDidSucceedNotification
                object:nil
                 queue:nil
            usingBlock:^(NSNotification *_Nonnull note) {
              if (weakSelf == nil | note.userInfo == nil) {
                return;
              }
              __strong typeof(self) strongSelf = weakSelf;

              FIRRemoteModel *model = note.userInfo[FIRModelDownloadUserInfoKeyRemoteModel];
              if ([model.name isEqualToString:@"your_remote_model"]) {
                // The model was downloaded and is available on the device
              }
            }];

[NSNotificationCenter.defaultCenter
    addObserverForName:FIRModelDownloadDidFailNotification
                object:nil
                 queue:nil
            usingBlock:^(NSNotification *_Nonnull note) {
              if (weakSelf == nil | note.userInfo == nil) {
                return;
              }
              __strong typeof(self) strongSelf = weakSelf;

              NSError *error = note.userInfo[FIRModelDownloadUserInfoKeyError];
            }];

Specify the model's input and output

Next, configure the model interpreter's input and output formats.

A TensorFlow Lite model takes as input and produces as output one or more multidimensional arrays. These arrays contain either byte, int, long, or float values. You must configure ML Kit with the number and dimensions ("shape") of the arrays your model uses.

If you don't know the shape and data type of your model's input and output, you can use the TensorFlow Lite Python interpreter to inspect your model. For example:

import tensorflow as tf

interpreter = tf.lite.Interpreter(model_path="my_model.tflite")
interpreter.allocate_tensors()

# Print input shape and type
print(interpreter.get_input_details()[0]['shape'])  # Example: [1 224 224 3]
print(interpreter.get_input_details()[0]['dtype'])  # Example: <class 'numpy.float32'>

# Print output shape and type
print(interpreter.get_output_details()[0]['shape'])  # Example: [1 1000]
print(interpreter.get_output_details()[0]['dtype'])  # Example: <class 'numpy.float32'>

After you determine the format of your model's input and output, configure your app's model interpreter by creating a ModelInputOutputOptions object.

For example, a floating-point image classification model might take as input an Nx224x224x3 array of Float values, representing a batch of N 224x224 three-channel (RGB) images, and produce as output a list of 1000 Float values, each representing the probability the image is a member of one of the 1000 categories the model predicts.

For such a model, you would configure the model interpreter's input and output as shown below:

Swift

let ioOptions = ModelInputOutputOptions()
do {
    try ioOptions.setInputFormat(index: 0, type: .float32, dimensions: [1, 224, 224, 3])
    try ioOptions.setOutputFormat(index: 0, type: .float32, dimensions: [1, 1000])
} catch let error as NSError {
    print("Failed to set input or output format with error: \(error.localizedDescription)")
}

Objective-C

FIRModelInputOutputOptions *ioOptions = [[FIRModelInputOutputOptions alloc] init];
NSError *error;
[ioOptions setInputFormatForIndex:0
                             type:FIRModelElementTypeFloat32
                       dimensions:@[@1, @224, @224, @3]
                            error:&error];
if (error != nil) { return; }
[ioOptions setOutputFormatForIndex:0
                              type:FIRModelElementTypeFloat32
                        dimensions:@[@1, @1000]
                             error:&error];
if (error != nil) { return; }

Perform inference on input data

Finally, to perform inference using the model, get your input data, perform any transformations on the data that might be necessary for your model, and build a Data object that contains the data.

For example, if your model processes images, and your model has input dimensions of [BATCH_SIZE, 224, 224, 3] floating-point values, you might have to scale the image's color values to a floating-point range as in the following example:

Swift

let image: CGImage = // Your input image
guard let context = CGContext(
  data: nil,
  width: image.width, height: image.height,
  bitsPerComponent: 8, bytesPerRow: image.width * 4,
  space: CGColorSpaceCreateDeviceRGB(),
  bitmapInfo: CGImageAlphaInfo.noneSkipFirst.rawValue
) else {
  return false
}

context.draw(image, in: CGRect(x: 0, y: 0, width: image.width, height: image.height))
guard let imageData = context.data else { return false }

let inputs = ModelInputs()
var inputData = Data()
do {
  for row in 0 ..< 224 {
    for col in 0 ..< 224 {
      let offset = 4 * (col * context.width + row)
      // (Ignore offset 0, the unused alpha channel)
      let red = imageData.load(fromByteOffset: offset+1, as: UInt8.self)
      let green = imageData.load(fromByteOffset: offset+2, as: UInt8.self)
      let blue = imageData.load(fromByteOffset: offset+3, as: UInt8.self)

      // Normalize channel values to [0.0, 1.0]. This requirement varies
      // by model. For example, some models might require values to be
      // normalized to the range [-1.0, 1.0] instead, and others might
      // require fixed-point values or the original bytes.
      var normalizedRed = Float32(red) / 255.0
      var normalizedGreen = Float32(green) / 255.0
      var normalizedBlue = Float32(blue) / 255.0

      // Append normalized values to Data object in RGB order.
      let elementSize = MemoryLayout.size(ofValue: normalizedRed)
      var bytes = [UInt8](repeating: 0, count: elementSize)
      memcpy(&bytes, &normalizedRed, elementSize)
      inputData.append(&bytes, count: elementSize)
      memcpy(&bytes, &normalizedGreen, elementSize)
      inputData.append(&bytes, count: elementSize)
      memcpy(&ammp;bytes, &normalizedBlue, elementSize)
      inputData.append(&bytes, count: elementSize)
    }
  }
  try inputs.addInput(inputData)
} catch let error {
  print("Failed to add input: \(error)")
}

Objective-C

CGImageRef image = // Your input image
long imageWidth = CGImageGetWidth(image);
long imageHeight = CGImageGetHeight(image);
CGContextRef context = CGBitmapContextCreate(nil,
                                             imageWidth, imageHeight,
                                             8,
                                             imageWidth * 4,
                                             CGColorSpaceCreateDeviceRGB(),
                                             kCGImageAlphaNoneSkipFirst);
CGContextDrawImage(context, CGRectMake(0, 0, imageWidth, imageHeight), image);
UInt8 *imageData = CGBitmapContextGetData(context);

FIRModelInputs *inputs = [[FIRModelInputs alloc] init];
NSMutableData *inputData = [[NSMutableData alloc] initWithCapacity:0];

for (int row = 0; row < 224; row++) {
  for (int col = 0; col < 224; col++) {
    long offset = 4 * (col * imageWidth + row);
    // Normalize channel values to [0.0, 1.0]. This requirement varies
    // by model. For example, some models might require values to be
    // normalized to the range [-1.0, 1.0] instead, and others might
    // require fixed-point values or the original bytes.
    // (Ignore offset 0, the unused alpha channel)
    Float32 red = imageData[offset+1] / 255.0f;
    Float32 green = imageData[offset+2] / 255.0f;
    Float32 blue = imageData[offset+3] / 255.0f;

    [inputData appendBytes:&red length:sizeof(red)];
    [inputData appendBytes:&green length:sizeof(green)];
    [inputData appendBytes:&blue length:sizeof(blue)];
  }
}

[inputs addInput:inputData error:&error];
if (error != nil) { return nil; }

After you prepare your model input (and after you confirm the model is available), pass the input and input/output options to your model interpreter's run(inputs:options:completion:) method.

Swift

interpreter.run(inputs: inputs, options: ioOptions) { outputs, error in
    guard error == nil, let outputs = outputs else { return }
    // Process outputs
    // ...
}

Objective-C

[interpreter runWithInputs:inputs
                   options:ioOptions
                completion:^(FIRModelOutputs * _Nullable outputs,
                             NSError * _Nullable error) {
  if (error != nil || outputs == nil) {
    return;
  }
  // Process outputs
  // ...
}];

You can get the output by calling the output(index:) method of the object that is returned. For example:

Swift

// Get first and only output of inference with a batch size of 1
let output = try? outputs.output(index: 0) as? [[NSNumber]]
let probabilities = output??[0]

Objective-C

// Get first and only output of inference with a batch size of 1
NSError *outputError;
NSArray *probabilites = [outputs outputAtIndex:0 error:&outputError][0];

How you use the output depends on the model you are using.

For example, if you are performing classification, as a next step, you might map the indexes of the result to the labels they represent. Suppose you had a text file with label strings for each of your model's categories; you could map the label strings to the output probabilities by doing something like the following:

Swift

guard let labelPath = Bundle.main.path(forResource: "retrained_labels", ofType: "txt") else { return }
let fileContents = try? String(contentsOfFile: labelPath)
guard let labels = fileContents?.components(separatedBy: "\n") else { return }

for i in 0 ..< labels.count {
  if let probability = probabilities?[i] {
    print("\(labels[i]): \(probability)")
  }
}

Objective-C

NSError *labelReadError = nil;
NSString *labelPath = [NSBundle.mainBundle pathForResource:@"retrained_labels"
                                                    ofType:@"txt"];
NSString *fileContents = [NSString stringWithContentsOfFile:labelPath
                                                   encoding:NSUTF8StringEncoding
                                                      error:&labelReadError];
if (labelReadError != nil || fileContents == NULL) { return; }
NSArray<NSString *> *labels = [fileContents componentsSeparatedByString:@"\n"];
for (int i = 0; i < labels.count; i++) {
    NSString *label = labels[i];
    NSNumber *probability = probabilites[i];
    NSLog(@"%@: %f", label, probability.floatValue);
}

Appendix: Model security

Regardless of how you make your TensorFlow Lite models available to ML Kit, ML Kit stores them in the standard serialized protobuf format in local storage.

In theory, this means that anybody can copy your model. However, in practice, most models are so application-specific and obfuscated by optimizations that the risk is similar to that of competitors disassembling and reusing your code. Nevertheless, you should be aware of this risk before you use a custom model in your app.